1 - (SYM 59) Re-evaluating Winter Depression Treatments: Unpacking the Unexpected Non-replication of CBT-SAD vs. Light Therapy Efficacy Amid the COVID-19 Pandemic

We conducted an NIMH-funded confirmatory efficacy trial comparing light therapy and cognitive-behavioral treatments for winter seasonal affective disorder (SAD). Adults with Major Depression, Recurrent with Seasonal Pattern (N=141) were randomized to 6-weeks of SAD-tailored group cognitive-behavioral therapy (CBT-SAD) or light therapy (LT) and followed one and two winters later. In a previous trial with the same design (N=177), post-treatment outcomes for CBT-SAD and LT were very similar, but CBT-SAD was associated with significantly fewer depression recurrences and less severe symptoms two winters later than LT (Rohan et al., 2015, 2016). The new study was funded in 2018, with a new cohort recruited/treated in each of the first 4 years. Data collection was in progress when the COVID-19 pandemic began.

Linear mixed models using weekly scores during treatment revealed significant Treatment X Time interactions on the Structured Interview Guide for the Hamilton Rating Scale for Depression-SAD Version (SIGH-SAD) and Beck Depression Inventory-II (BDI-II). LT worked faster than CBT-SAD, with lower depression severity most weeks, but no significant difference at post-treatment. There were significantly more remissions at post-treatment in LT than in CBT-SAD using SIGH-SAD (60.3% vs. 41.2%) and BDI-II (75.0% vs. 58.0%) criteria. In the primary intent-to-treat analysis based on multiple imputation of missing outcome scores, the treatments did not differ significantly on followup outcomes, with very similar proportions of recurrences at first (17.4% CBT-SAD vs. 23.8% LT) and second (31.8% CBT-SAD vs. 28.9% LT) winter followup.

These results differ strikingly from the prior study. The COVID-19 pandemic is an obvious difference between the two. To examine whether the pandemic affected treatment efficacy in the current study, we conducted exploratory analyses by cohort. Results suggest: (1) COVID had noticeable effects on CBT-SAD but not on LT. (2) CBT-SAD outcomes were best for cohort 1, which received treatment and completed both followups before the pandemic. (3) The treatment difference in post-treatment remission rates was ns prior to the onset of COVID-19 and became significant after. (4) Compared to CBT-SAD participants treated during COVID (cohorts 3 and 4), those treated before COVID (cohorts 1 and 2) had fewer recurrences at first winter followup (7.3% vs. 32.1%). It is possible that the necessary switch from in-person to teleconference CBT-SAD during COVID-19 restrictions negatively affected its efficacy and durability.